Identificador persistente para citar o vincular este elemento: http://hdl.handle.net/10553/106990
Campo DC Valoridioma
dc.contributor.authorCastillo Bolado, David Alejandroen_US
dc.contributor.authorGuerra-Artal, Cayetanoen_US
dc.contributor.authorHernÁndez-Tejera, Marioen_US
dc.date.accessioned2021-04-26T08:28:04Z-
dc.date.available2021-04-26T08:28:04Z-
dc.date.issued2021en_US
dc.identifier.issn0893-6080en_US
dc.identifier.otherScopus-
dc.identifier.otherWoS-
dc.identifier.urihttp://hdl.handle.net/10553/106990-
dc.description.abstractMonolithic neural networks and end-to-end training have become the dominating trend in the field of deep learning, but the steady increase in complexity and training costs has raised concerns about the effectiveness and efficiency of this approach. We propose modular training as an alternative strategy for building modular neural networks by composing neural modules that can be trained independently and then kept for future use. We analyse the requirements and challenges regarding modularity and compositionality and, with that information in hand, we provide a detailed design and implementation guideline. We show experimental results of applying this modular approach to a Visual Question Answering (VQA) task parting from a previously published modular network and we evaluate its impact on the final performance, with respect to a baseline trained end-to-end. We also perform compositionality tests on CLEVR.en_US
dc.languageengen_US
dc.relation.ispartofNeural Networksen_US
dc.sourceNeural Networks [ISSN 0893-6080], v. 139, p. 294-304, (Julio 2021)en_US
dc.subject120304 Inteligencia artificialen_US
dc.subject.otherCompositionalityen_US
dc.subject.otherLearning by roleen_US
dc.subject.otherMethodologyen_US
dc.subject.otherModular trainingen_US
dc.subject.otherModularityen_US
dc.subject.otherVisual question answeringen_US
dc.titleDesign and independent training of composable and reusable neural modulesen_US
dc.typeinfo:eu-repo/semantics/Articleen_US
dc.typeArticleen_US
dc.identifier.doi10.1016/j.neunet.2021.03.034en_US
dc.identifier.scopus85104133190-
dc.identifier.isi000652682000011-
dc.contributor.authorscopusid57208508176-
dc.contributor.authorscopusid57217392692-
dc.contributor.authorscopusid55966875800-
dc.identifier.eissn1879-2782-
dc.description.lastpage304en_US
dc.description.firstpage294en_US
dc.relation.volume139en_US
dc.investigacionIngeniería y Arquitecturaen_US
dc.type2Artículoen_US
dc.contributor.daisngid44219620-
dc.contributor.daisngid5140885-
dc.contributor.daisngid2188888-
dc.description.numberofpages11en_US
dc.utils.revisionen_US
dc.contributor.wosstandardWOS:Castillo-Bolado, D-
dc.contributor.wosstandardWOS:Guerra-Artal, C-
dc.contributor.wosstandardWOS:Hernandez-Tejera, M-
dc.date.coverdateJulio 2021en_US
dc.identifier.ulpgcen_US
dc.contributor.buulpgcBU-INFen_US
dc.description.sjr2,781
dc.description.jcr9,657
dc.description.sjrqQ1
dc.description.jcrqQ1
dc.description.scieSCIE
dc.description.miaricds11,0
dc.description.erihplusERIH PLUS
item.fulltextCon texto completo-
item.grantfulltextopen-
crisitem.author.deptGIR SIANI: Inteligencia Artificial, Redes Neuronales, Aprendizaje Automático e Ingeniería de Datos-
crisitem.author.deptIU Sistemas Inteligentes y Aplicaciones Numéricas-
crisitem.author.deptDepartamento de Informática y Sistemas-
crisitem.author.deptGIR SIANI: Inteligencia Artificial, Redes Neuronales, Aprendizaje Automático e Ingeniería de Datos-
crisitem.author.deptIU Sistemas Inteligentes y Aplicaciones Numéricas-
crisitem.author.deptDepartamento de Informática y Sistemas-
crisitem.author.orcid0000-0003-1381-2262-
crisitem.author.orcid0000-0001-9717-8048-
crisitem.author.parentorgIU Sistemas Inteligentes y Aplicaciones Numéricas-
crisitem.author.parentorgIU Sistemas Inteligentes y Aplicaciones Numéricas-
crisitem.author.fullNameCastillo Bolado, David Alejandro-
crisitem.author.fullNameGuerra Artal, Cayetano-
crisitem.author.fullNameHernández Tejera, Francisco Mario-
Colección:Artículos
Adobe PDF (853,53 kB)
Vista resumida

Visitas

196
actualizado el 31-ago-2024

Descargas

397
actualizado el 31-ago-2024

Google ScholarTM

Verifica

Altmetric


Comparte



Exporta metadatos



Los elementos en ULPGC accedaCRIS están protegidos por derechos de autor con todos los derechos reservados, a menos que se indique lo contrario.